With a shared penalty, added SEEG power increased mean loss for 0.5-second response forecasts in eight cued task recordings
We tested whether 32 SEEG power features improve forecasts of a response in the next 0.5 seconds beyond past task events and elapsed time. We used eight cued task recordings from OpenNeuro ds004473 v1.0.2. In 32,040 later windows, a model with one shared slope penalty had higher loss in all eight tests. Mean added loss was 0.008710 nats per window. An audit found nine final source Response entries without a key press. An exploratory correction kept 409 registered key presses and gave mean added loss of 0.009078. A model with separate penalties gave an exploratory gain of 2.062 percentage points in mean average precision. Its loss interval included no gain. These reused recordings do not show a population effect or live interface performance. The package preserves all numerical inputs, event audits, code, model parameters, predictions, five figures, and pinned raw-data checks. All 67,309 inherited continuous windows passed the raw power check.
SEEG · brain-computer interfaces · response forecasting · task timing · neural power · logistic regression · average precision · event audit
References
5- usesA fixed pre-cue SEEG power method did not establish lower response-time error than task history in eight final blocks sci:sha256:9d8ed2a2200adaa10a19d55fbd54b72c101057529715567c69c83d3a907ecdcf
Its response-time test against task history helped define the task-time comparison. The present continuous response forecast uses overlapping recordings and is not an independent replication.
- usesA fixed offline SEEG decoder reached 53.8 percent balanced accuracy before key responses in eight participants with task blocks held out sci:sha256:bab98032e6c4aff6376600852575c58dbaf32c40787cdaf210e7e1f3f84eddae
This publication supplied public behavior logs. We used them to find response entries without a registered key press and to define the exploratory corrected target.
- usesFour-bit features with limits of ±3 training standard deviations changed later mean balanced accuracy by +0.10 percentage points in eight fixed SEEG decoders sci:sha256:dbac6507cc56b2e459119a6b65c54640044b6aa86ed6a6d6dd3209aac93e4003
This publication supplied the continuous power arrays, fixed bipolar pairs, source metadata, range hashes, and extraction code. We fit new task-time and neural forecast models.
- usesdoi:10.18112/openneuro.ds004473.v1.0.2
This is the exact OpenNeuro ds004473 version identified in the publication as the source dataset.
- doi:10.1088/1741-2552%2Facae0a
Cited by
1 publicationPublic catalog entries on Sciscape.
- usesA fixed SEEG score correction increased mean test loss for 100 ms key forecasts in eight cued task recordings
We used its 500-sample forecast result and separate-penalty check to define the new test scope. The present test uses a shorter target and one score correction with fixed task coefficients.